基于加权马尔柯夫链的汛期降水丰枯预测

Forecasting Precipitation During Flood Period Based on the Model of Markov Chain with Weights

依据榆林市秃尾河1976-2000年汛期降水资料,应用均值标准差法建立降水丰枯的分级标准,将汛期降水量分为枯、偏枯、平、偏丰、丰等5个水平年。然后针对降水量为相依随机变量的特点,采用以规范化的各阶自相关系数为权重,运用加权马尔柯夫链模型来预测和分析未来年份汛期降水量的丰枯状况,结果显示该方法是有效可行的。最后运用遍历定理进行各种滞时的马尔柯夫链特征分析,对降水序列做出了定量的描述,预测未来秃尾河流域汛期降水量出现丰水年的几率最大,为39.8%。通过对汛期降水状态的预测,可以更好的为流域防洪抗汛和水库的合理调度提供有效信息。

Mean-standard deviation method is used to establish the classification standard of precipitation,which divides the precipitation into dry year,relatively dry year,normal year,relatively wet year,and wet year,on the basis of the flood precipitation data from 1976 to 2000 for the Tuwei River in Yulin City.Then this paper presents a method called Markov Chain with weights to predict the future precipitation of flood period by regarding the standardized self-coefficients as weights based on the special characteristics of precipitation.The results show that this method is effective and feasible.Finally ergodic theorem is used to analyze various hysteresis characteristics of the Markov Chain,and make a quantitative analysis of the precipitation sequence that precipitation for wet years is most likely to occur.It can provide effective information for flood control and reasonable dispatch in reservoirs of the river basin based on the flood forecasting precipitation.